Quick check

Six things AI needs to work at scale

Move each slider to what fits your organisation today. You'll see what it takes, and get a summary you can use when you talk to AI vendors.

    Details per area

    1

    What AI should do

    Capability

    What do we want AI to help us with?

    Level 1 | Ask and write

    Your situation

    We want help writing, summarising and answering questions about documents.

    What it takes

    A standard AI assistant. Most will do: Copilot, ChatGPT, Claude, Gemini, or open models such as gpt-oss, Gemma, Qwen and Mistral.

    Level 2 | Listen and find

    Your situation

    We also want meetings transcribed, texts translated and answers found in our own documents.

    What it takes

    Add speech-to-text and a search that looks in your own documents. Open models handle most of this.

    Level 3 | Fill in our systems

    Your situation

    We want AI to fetch and enter information directly in the systems we work in.

    What it takes

    Strong AI models that can work with other software, plus links (connectors) to your systems.

    Level 4 | Do tasks on its own

    Your situation

    We want AI to complete tasks with several steps on its own, with a person approving when needed.

    What it takes

    The most capable AI models, a way to coordinate multi-step tasks, human sign-off on important steps, and custom links to your systems.

    2

    Control of your data

    Sovereignty

    How much can outside AI companies see of our data?

    Level 1 | Anywhere is fine

    Your situation

    Our data is public or not sensitive. It doesn't matter much where it is processed.

    What it takes

    Any well-known AI service with a standard business contract and a data processing agreement.

    Level 2 | Kept in the EU

    Your situation

    We share internal data and some personal data covered by GDPR.

    What it takes

    Data stored and processed in the EU, a written promise that your data is not used to train AI, and a data processing agreement.

    Level 3 | Under European control

    Your situation

    We share sensitive or business-critical data, and want to limit access from countries outside the EU.

    What it takes

    A European or sovereign cloud running AI models hosted in the EU. Non-EU AI services only for tasks without sensitive data.

    Level 4 | Never leaves our walls

    Your situation

    Our data is classified or critical to society. It must not leave our own environment.

    What it takes

    AI running on your own servers, possibly cut off from the internet, with models you run yourself and your own encryption keys.

    3

    Rules for use

    Permission

    What may we use AI for, and how do we make sure people stick to it?

    Level 1 · Common sense

    Your situation

    Staff can use AI freely within general guidelines, for low-risk tasks.

    What it takes

    An AI policy, training for staff, and the settings built into the AI tool.

    Level 2 · Approved tools

    Your situation

    Only approved tools and tasks are allowed. Some departments have limits.

    What it takes

    Central approval of tasks, one company login, and access based on role.

    Level 3 · Rules per task

    Your situation

    Rules differ from task to task, and how sensitive the data is decides which AI may be used.

    What it takes

    Rules set up per task (who, which data, which AI), automatic checks on what goes in and comes out, and a list of all approved AI uses.

    Level 4 · Strictly enforced

    Your situation

    AI supports decisions about citizens, patients or customers, and the EU AI Act or sector rules apply.

    What it takes

    A system that enforces the rules automatically: risk assessment, approval steps, ongoing checks and human oversight.

    4

    Proof and documentation

    Audit

    How much must we be able to prove about how AI is used?

    Level 1 · Basic overview

    Your situation

    No formal requirements. Knowing how much AI is used is enough.

    What it takes

    The usage overview built into the AI tool.

    Level 2 · Regular reports

    Your situation

    We report internally on who uses what, and what it costs.

    What it takes

    Central records of use per person and tool, with regular reports.

    Level 3 · Ready for audit

    Your situation

    We must show auditors and our data protection officer what AI is used for, the risks and the controls.

    What it takes

    A list of all AI uses, records of questions and answers for selected tasks, rules for how long records are kept, and written risk assessments.

    Level 4 · Every step traceable

    Your situation

    Regulators, courts or citizens must be able to see exactly how AI was used in each case.

    What it takes

    Complete records that can't be changed (questions, answers, AI version and decisions), easy export, and ongoing testing and monitoring.

    5

    Connection to your systems

    Integration

    Which of our systems and data does AI need to work with?

    Level 1 · Copy and paste

    Your situation

    We copy text into AI or upload files ourselves.

    What it takes

    A ready-to-use AI assistant. No connection to other systems.

    Level 2 · Reads our documents

    Your situation

    AI must be able to read our documents and intranet.

    What it takes

    Standard links to your document storage, where people only see what they already have access to.

    Level 3 · Updates our systems

    Your situation

    AI must read and update our case, finance or customer systems.

    What it takes

    Custom links to those systems that respect existing access rights, plus budget for setting them up.

    Level 4 · Built into our processes

    Your situation

    AI runs whole processes across several systems, also without a person involved.

    What it takes

    A setup to connect, coordinate, monitor and test it all. Plan it as a long-term programme, not a one-off project.

    6

    Cost at scale

    Economics

    What will it cost when many people use it?

    Level 1 · Try it out

    Your situation

    We have a few dozen users on a fixed budget.

    What it takes

    A price per user or pay per use. Low and predictable.

    Level 2 · A few departments

    Your situation

    We have hundreds of users across a few departments.

    What it takes

    Company-wide licences, costs followed per department, and a comparison of paying per user versus per use.

    Level 3 · Whole organisation

    Your situation

    We have thousands of users, plus AI running inside our systems.

    What it takes

    Pay per use, with costs split per task and department, spending limits, and cheaper AI for simple tasks.

    Level 4 · Core operations

    Your situation

    AI is part of our core operations and runs many automated tasks.

    What it takes

    Your own or reserved computing capacity for a predictable price per task, and ongoing cost monitoring.

    Your summary

    A conclusion based on all six choices together. Use it as your list of requirements when you talk to AI vendors.

    In short

    Your profile points to an AI platform that runs entirely in your own environment, with rules and documentation built in from day one.

    Your profile points to AI that runs entirely in your own environment, on models you operate yourselves.

    Your profile points to a European-hosted AI platform with rules and documentation built in, rather than a standard AI subscription.

    Your profile points to AI hosted under European control, with open models for anything sensitive.

    A mainstream AI service can work for you, but only with a governance layer on top that enforces your rules and documents how AI is used.

    A mainstream AI service that keeps your data in the EU fits your needs. Most of the effort will go into connecting it to your systems.

    A standard AI assistant covers your needs today. Start simple, and take this check again as use grows.

    AI models

    Only open models you run on your own servers, such as gpt-oss-120b, Qwen, Mistral or Gemma. Choose the largest your hardware allows, because system work and multi-step tasks need strong models. ChatGPT, Claude and Gemini are ruled out.

    Only open models you run on your own servers, such as gpt-oss, Qwen, Mistral or Gemma. Mid-size models are enough for writing, summaries and search. ChatGPT, Claude and Gemini are ruled out.

    Large open models hosted in the EU, such as gpt-oss-120b, Qwen or Mistral Large, for everything involving sensitive data. ChatGPT, Claude or Gemini only for tasks without sensitive data, routed through a gateway that enforces this.

    Open models hosted in the EU, such as gpt-oss, Qwen, Mistral or Gemma, cover your tasks. Services from outside the EU only for tasks without sensitive data.

    The strongest models, such as ChatGPT, Claude or Gemini, on plans that keep your data in the EU. Check each vendor, as EU data residency isn't offered for every product.

    Any mainstream assistant on a plan that keeps your data in the EU and doesn't use it for training: Copilot, ChatGPT, Gemini or an EU-hosted open model.

    The strongest models, such as ChatGPT, Claude or Gemini, give the best results for system work and agents.

    Any mainstream assistant will do: Copilot, ChatGPT, Claude, Gemini, or an open model.

    Hosting and data

    Your own servers or an isolated data centre, with your own encryption keys. Plan for GPU hardware and people to run it, or a partner who operates it inside your walls.

    A European-owned or sovereign cloud, so no company outside the EU can be forced to hand over your data. Ask vendors who owns the infrastructure and which country's law applies.

    Data stored and processed in the EU, a data processing agreement, and a written promise that your data is not used for training.

    A standard business contract and a data processing agreement are enough.

    Rules and documentation

    You need a governance platform that enforces your rules automatically and keeps records that can't be changed of every question, answer, AI version and decision. The admin settings built into AI tools are not enough at this level.

    You need rules per task and records good enough for an audit: a list of every AI use, automatic checks on what goes in and comes out, and logs for selected tasks.

    You need rules per task with automatic checks, and a list of approved AI uses. Your documentation needs are lighter, so standard usage logs will do.

    Your rules are simple, but you must be able to show auditors how AI is used. Keep a list of AI uses, and log questions and answers for selected tasks.

    One company login, a list of approved tools and regular usage reports cover your needs.

    An AI policy, staff training and the AI tool's built-in settings are enough.

    Connection to your systems

    Plan for an integration programme. AI that works across several systems needs coordination, monitoring, test environments and human sign-off on important steps.

    Budget for custom links to your case, finance or customer systems, and make sure AI only sees what each user is allowed to see.

    Standard links to your document storage and intranet, where people only see what they already have access to.

    No integration needed. People work directly in the AI tool.

    Cost

    Running AI on your own servers has a high fixed starting cost for a small number of users. Consider a shared or partner-operated setup until use grows.

    At your scale, dedicated capacity in Europe or in your own data centre gives a predictable price per task. Compare it with paying per use before you commit.

    Reserve or own computing capacity for a predictable price per task, and follow cost per task as a key figure.

    Pay per use, with budgets and spending limits per department, and send simple tasks to cheaper models.

    Licences for the whole organisation, with cost followed per department. Compare paying per user with paying per use.

    A price per user or pay per use keeps the cost low and predictable.

    Watch out for

    The most capable AI models can't run on your own servers. Expect lower quality on the hardest multi-step tasks, or allow tasks without sensitive data to use external services.

    Agents and system work on open models hosted in the EU is possible, but quality varies. Test on your own tasks before you commit.

    Rules you can't document are hard to prove. Even simple logging helps when someone asks how AI was used.

    AI that completes tasks on its own needs clear rules for what it may do, even if other AI use is relaxed.

    AI that updates your systems needs more than common sense. Decide which tasks it may do before you connect it.

    Thousands of users with only general guidelines makes unapproved AI use hard to control. Consider approved tools and one company login.

    Next step

    Share this summary with the vendors you are considering, and ask each one to show how they meet every point. Points they can't meet show where you need a partner or a different solution.

    Book a demo